On Mon, 2014-02-24 at 22:31 +0200, Luca Coelho wrote:
> > > + /* TODO: in cfg80211_mlme_auth() we used to check if the
> > > + * channel can be used *before* this function was called. Do
> > > + * we still need to do it or can we rely on the combinations
> > > + * check that will happen later, in
> > > + * ieee80211_vif_use_channel()?
> > > + */
> >
> > We'll do it a few lines down in ieee80211_prep_connection(), so I don't
> > see why we'd need to do it earlier?
>
> Yes, you're right. I just wondered whether we should avoid doing a
> deauth before we checked the combinations when we are associated (just
> before we call ieee80211_prep_connection()).
>
> I will remove the TODO if you think we shouldn't worry about this.
I wouldn't really be worried about this. If it turns out to be a problem
we can always fix it later, but I doubt it.
> > > @@ -406,6 +406,9 @@ cfg80211_can_change_interface(struct cfg80211_registered_device *rdev,
> > > struct wireless_dev *wdev,
> > > enum nl80211_iftype iftype)
> > > {
> > > + /* TODO: For this function, we'll probably need to keep some
> > > + * kind of interface combination check in cfg80211...
> > > + */
> > > return cfg80211_can_use_iftype_chan(rdev, wdev, iftype, NULL,
> > > CHAN_MODE_UNDEFINED, 0);
> > > }
> >
> > We could always jsut request the change from the driver/mac80211.
>
>
> This is also called by cfg80211_can_add_interface() which is called in
> PRE_UP and start_p2p_device.
>
> And in the cfg80211_change_interface() function, we do stuff like
> stopping an AP or disconnecting a client. I definitely think we should
> check if the change is possible before taking these actions.
Ok.
> Do you mean to add an rdev_op to handle this?
No I was thinking of just (re)moving it.
johannes
--
To unsubscribe from this list: send the line "unsubscribe linux-wireless" in
the body of a message to majordomo@xxxxxxxxxxxxxxx
More majordomo info at http://vger.kernel.org/majordomo-info.html
